g(dZddlZddlZddlddlmZddlmZddlddl m Z Gdde Z Gd d eZ Gd d eZid ddddddddddddddddddd d!d"d#d$d%d&d'd(d)d*d+d,d-d.id/d0d1d2d3d4d5d6d7d8d9d:d;d<d=d>d?d@dAdBdCdDdEdFdGdHdIdJdKdLdMdNdOdNidPdQdRdSdTdUdVdWdXdYdZd[d\d]d^d_d`dadbdcdddedfdgdhdidjdkdldmdndodpdqidrdsdtdudvdwdxdydzd{d|d}d~dddddddddddddddddddddiddddddddddddddddddddddddddddddddddddddddddddddddddŜZy)a Corpus reader for the York-Toronto-Helsinki Parsed Corpus of Old English Prose (YCOE), a 1.5 million word syntactically-annotated corpus of Old English prose texts. The corpus is distributed by the Oxford Text Archive: http://www.ota.ahds.ac.uk/ It is not included with NLTK. The YCOE corpus is divided into 100 files, each representing an Old English prose text. Tags used within each text complies to the YCOE standard: https://www-users.york.ac.uk/~lang22/YCOE/YcoeHome.htm N)*)BracketParseCorpusReader)TaggedCorpusReader)RegexpTokenizercfeZdZdZddZddZddZdZddZddZ dd Z dd Z dd Z dd Z dd Zy)YCOECorpusReaderz Corpus reader for the York-Toronto-Helsinki Parsed Corpus of Old English Prose (YCOE), a 1.5 million word syntactically-annotated corpus of Old English prose texts. ctj||g|t|jj ddd||_t |jj ddd|_|j jDchc]}|dd }}|jjDchc]}|dd c}|k7r tdt|Dcgc]}d |z c}|Dcgc]}d |z c}z}tj||||t||_ ycc}wcc}wcc}wcc}w) Npsdz.*.psd)encodingpos.posz5Items in "psd" and "pos" subdirectories do not match.%s.psd%s.pos) CorpusReader__init__YCOEParseCorpusReaderrootjoin _psd_readerYCOETaggedCorpusReader _pos_readerfileids ValueErrorsorted _documents)selfrr f documentsdocrs L/var/www/openai/venv/lib/python3.12/site-packages/nltk/corpus/reader/ycoe.pyrzYCOECorpusReader.__init__%s.dD"h70 IINN5 !4( 2$))..2GvV&*%5%5%=%=%?@%?QsV%? @ ,,446 76qAcrF6 79 DWX X'0 1yX^y 1)23#x#~3 4  dD'8< +A 7 23s D-7 D2 D7/ D< Nc| |jSt|tr|g}|D]}||jvst d|zt |Dchc]}|dd c}Scc}w)z Return a list of document identifiers for all documents in this corpus, or for the documents with the given file(s) if specified. NzFile id %s not foundr)r isinstancestr_fileidsKeyErrorr)rrrs r"r zYCOECorpusReader.documents9sp ??? " gs #iGA %5?@@w/w!q"vw/00/s A!c | |jSt|tr|g}tt |Dcgc]}d|z c}|Dcgc]}d|z c}zScc}wcc}w)z Return a list of file identifiers for the files that make up this corpus, or that store the given document(s) if specified. rr)r&r$r%rset)rr r!s r"rzYCOECorpusReader.fileidsIsq  ==  3 '" I +459CC95-67Yc8c>Y78   57s A  A c| |j}nHt|tr|g}|D]0}||jvs|dddvr tdtd|z|Dcgc] }|d| c}Scc}w)z Helper that selects the appropriate fileids for a given set of documents from a given subcorpus (pos or psd). Nr)rr zvExpected a document identifier, not a file identifier. (Use corpus.documents() to get a list of document identifiers.z Document identifier %s not found.)rr$r%r)rr subcorpusdocumentds r" _getfileidszYCOECorpusReader._getfileidsYs  I)S)&K %4??2}(88(> ))Kh)VWW&-66Iq1#Qyk"I666sA-cX|jj|j|dSNr )rwordsr/rr s r"r2zYCOECorpusReader.wordsp&%%d&6&6y%&HIIcX|jj|j|dSr1)rsentsr/r3s r"r7zYCOECorpusReader.sentssr4r5cX|jj|j|dSr1)rparasr/r3s r"r9zYCOECorpusReader.parasvr4r5cX|jj|j|dSr1)r tagged_wordsr/r3s r"r;zYCOECorpusReader.tagged_wordsy&,,T-=-=i-OPPr5cX|jj|j|dSr1)r tagged_sentsr/r3s r"r>zYCOECorpusReader.tagged_sents|r<r5cX|jj|j|dSr1)r tagged_parasr/r3s r"r@zYCOECorpusReader.tagged_parasr<r5cX|jj|j|dS)Nr )r parsed_sentsr/r3s r"rBzYCOECorpusReader.parsed_sentsr<r5utf8)N)__name__ __module__ __qualname____doc__rr rr/r2r7r9r;r>r@rBr5r"rrsE ,(1  7.JJJQQQQr5rceZdZdZdZy)rzrSpecialized version of the standard bracket parse corpus reader that strips out (CODE ...) and (ID ...) nodes.ctjdd|}tjd|rytj||S)Nz(?u)\((CODE|ID)[^\)]*\)z\s*\(\s*\)\s*$)resubmatchr_parse)rts r"rPzYCOEParseCorpusReader._parses: FF-r1 5 88%q )'..tQ77r5N)rErFrGrHrPrIr5r"rrs 68r5rceZdZddZy)rcVd}t|d}tj|||d|y)Nz+(?u)(?<=/\.)\s+|\s*\S*_CODE\s*|\s*\S*_ID\s*T)gaps_)sepsent_tokenizer)rrr)rritemsr gaps_rerWs r"rzYCOETaggedCorpusReader.__init__s,@(t<## $3~ r5NrC)rErFrGrrIr5r"rrs r5rz coadrian.o34zAdrian and Ritheusz coaelhom.o3uÆlfric, Supplemental Homiliesz coaelive.o3uÆlfric's Lives of SaintscoalcuinzAlcuin De virtutibus et vitiisz coalex.o23zAlexander's Letter to Aristotlez coapollo.o3zApollonius of Tyrecoaugust Augustinez cobede.o2z$Bede's History of the English Churchz cobenrul.o3zBenedictine Rulez coblick.o23zBlickling Homiliesz coboeth.o2z#Boethius' Consolation of Philosophyz cobyrhtf.o3zByrhtferth's Manual cocanedgDzCanons of Edgar (D) cocanedgXzCanons of Edgar (X)z cocathom1.o3uÆlfric's Catholic Homilies Iz cocathom2.o3uÆlfric's Catholic Homilies IIz cochad.o24z Saint ChadcochdrulzChrodegang of Metz, Rule cochristophzSaint Christopherz cochronA.o23zAnglo-Saxon Chronicle AcochronCzAnglo-Saxon Chronicle CcochronDzAnglo-Saxon Chronicle Dz cochronE.o34zAnglo-Saxon Chronicle Ez cocura.o2zCura PastoraliscocuraCzCura Pastoralis (Cotton)z codicts.o34z Dicts of Catoz codocu1.o1zDocuments 1 (O1)z codocu2.o12zDocuments 2 (O1/O2)z codocu2.o2zDocuments 2 (O2)z codocu3.o23zDocuments 3 (O2/O3)z codocu3.o3zDocuments 3 (O3)z codocu4.o24zDocuments 4 (O2/O4)coeluc1z Honorius of Autun, Elucidarium 1coeluc2z coepigen.o3uÆlfric's Epilogue to GenesiscoeuphrzSaint Euphrosynecoeustz Saint Eustace and his companions coexodusPz Exodus (P) cogenesiCz Genesis (C)z cogregdC.o24zGregory's Dialogues (C)z cogregdH.o23zGregory's Dialogues (H)coherbarzPseudo-Apuleius, Herbariumz coinspolD.o34z"Wulfstan's Institute of Polity (D) coinspolXz"Wulfstan's Institute of Polity (X)cojamesz Saint Jamesz colacnu.o23Lacnungaz colaece.o2 Leechdomsz colaw1cn.o3z Laws, Cnut Iz colaw2cn.o3z Laws, Cnut IIz colaw5atr.o3uLaws, Æthelred Vz colaw6atr.o3uLaws, Æthelred VIz colawaf.o2z Laws, Alfredz colawafint.o2zAlfred's Introduction to Lawsz colawger.o34z Laws, Gerefaz colawine.ox2z Laws, Inezcolawnorthu.o3zNorthumbra Preosta Laguz colawwllad.o4zLaws, William I, Ladz coleofri.o4Leofricz colsigef.o3uÆlfric's Letter to Sigefyrth colsigewBu!Ælfric's Letter to Sigeweard (B)z colsigewZ.o34u!Ælfric's Letter to Sigeweard (Z)colwgeatuÆlfric's Letter to Wulfgeat colwsigeTu Ælfric's Letter to Wulfsige (T)zcolwsigeXa.o34u!Ælfric's Letter to Wulfsige (Xa)z colwstan1.o3uÆlfric's Letter to Wulfstan Iz colwstan2.o3uÆlfric's Letter to Wulfstan IIz comargaC.o34zSaint Margaret (C)comargaTzSaint Margaret (T)comart1zMartyrology, Icomart2zMartyrology, IIz comart3.o23zMartyrology, IIIz comarvel.o23zMarvels of the Eastcomaryz Mary of Egyptconeotz Saint NeotconicodAzGospel of Nicodemus (A)conicodCzGospel of Nicodemus (C)conicodDzGospel of Nicodemus (D)conicodEzGospel of Nicodemus (E)z coorosiu.o2Orosiusz cootest.o3 Heptateuchzcoprefcath1.o3u(Ælfric's Preface to Catholic Homilies Izcoprefcath2.o3u)Ælfric's Preface to Catholic Homilies IIz coprefcura.o2zPreface to the Cura Pastoralisz coprefgen.o3uÆlfric's Preface to Genesiszcopreflives.o3u$Ælfric's Preface to Lives of Saintsz"Preface to Augustine's Soliloquiesz*Pseudo-Apuleius, Medicina de quadrupedibuszHistory of the Holy Rood-TreezSeven SleeperszSt. Augustine's SoliloquieszSolomon and Saturn IzSolomon and Saturn IIuÆlfric's De Temporibus AnnizVercelli HomilieszVercelli Homilies (E)zVercelli Homilies (L)zSaint Vincent (Bodley 343)zVindicta SalvatoriszWest-Saxon GospelszWulfstan's Homilies) coprefsoliloz coquadru.o23corood cosevenslcosoliloz cosolsat1.o4 cosolsat2z cotempo.o3coverhom coverhomE coverhomLcovinceBcovinsalz cowsgosp.o3z cowulf.o34)rHosrMnltk.corpus.reader.api nltk.corpus.reader.bracket_parsernltk.corpus.reader.taggedrnltk.corpus.reader.util nltk.tokenizerrrrrr rIr5r"rs  $E8%)eQ|eQP848 / e (e 3e .e 0 e  3 e  ' e  e 7e %e 'e 7e (e &e &e 3e  4!e ",#e $*%e &&'e (-)e *)+e ,)-e .-/e 0"1e 2)3e 4?5e 6$7e 8(9e :$;e <(=e >$?e @(Ae B1Ce D1Ee F2Ge H!Ie J 0Ke LMe NOe P-Qe R-Se T,Ue V9We X5Ye Z}[e \:]e ^+_e `>ae b?ce d'ee f(ge h.ie j4ke lNme nKoe p/qe r+se t9ue v2we x4ye z8{e |.}e ~3e @9Ae B4Ce D5Ee F(Ge H$Ie JKe L Me N%Oe P)Qe R oSe T lUe V)We X)Ye Z)[e \)]e ^9_e `,ae b@ce dAee f5ge h2ie j